Image Recognition and Noise Filtering on Flooded-Car Study (val)
95.4PrecisionReCCur
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| ReCCur2026.01 | 95.4 | 97.2 | 96.3 | 95.6 | 99 | |
| GPT-5Retrieval-Augmented=True2026.01 | 90.2 | 88.5 | 87.3 | 85.2 | 87.9 | |
| GPT-5Protocol=Instruction QA2026.01 | 89.2 | 86.3 | 84.6 | 83.9 | 88.6 | |
| Vo et al.2026.01 | 88.6 | 57.6 | 68.5 | 77.7 | 81.5 | |
| Qwen-VL 2.5Protocol=Instruction QA2026.01 | 84.6 | 79.9 | 78.3 | 78.4 | 83.1 | |
| K-means2026.01 | 81.1 | 68.3 | 69.1 | 49.7 | 100 | |
| KNN2026.01 | 78.8 | 89.8 | 83.2 | 78.5 | 98.2 | |
| CaSED2026.01 | 77.4 | 58.3 | 48.1 | 100 | 17.3 | |
| SigLIP22026.01 | 74.8 | 47.4 | 39.3 | 41.2 | 100 | |
| YOLO-World v22026.01 | 25.3 | 12.9 | 9.4 | 96.8 | 3.2 |